ICMP polling via proxy not working

[bikeprata]

Hello everyone!

I'm having trouble on using a external agent to do ICMP polling on version 2.0-M4.

I created a new zone, sellected a working agent to monitor that zone and put a working node there (node stats were UP and low latency response). After that, the node status went down and polling requests from server returns that the node is unreacheble.

This is very odd because I checked via tcpdump that the agent is receiving the polling request and is sending a ICMP packet to the node. The node is answering ok with another ICMP packet that I could see on tcpdump.

It seems that somehow the agent isn't sending the info that the node is UP back to server.

Have you experienced something like that? Is there any chance that I'm doing something wrong?

Thanks in advance for your help!

[d-ray]

I have the same issue... I post the issue already

[bikeprata]

Thanks, d-ray!

[bikeprata]

IMPORTANT UPDATE:

[06-May-2015 17:33:44.897] [DEBUG] StatusPoll(10.5.1.200): starting child object poll
[06-May-2015 17:33:44.913] [DEBUG] StatusPoll(10.5.1.200): polling interface 124 [unknown]
[06-May-2015 17:33:44.913] [DEBUG] Interface::StatusPoll(124,unknown): ping via proxy [114]
[06-May-2015 17:33:44.913] [DEBUG] Interface::StatusPoll(124,unknown): proxy node found: nodebmon
[06-May-2015 17:33:44.961] [DEBUG] Interface::StatusPoll(124,unknown): proxy response: "30"
[06-May-2015 17:33:44.962] [DEBUG] Interface::StatusPoll(124,unknown): newStatus=5 oldStatus=5 pollCount=6 requiredPolls=1
[06-May-2015 17:33:44.962] [DEBUG] StatusPoll(10.5.1.200): finished child object poll
[06-May-2015 17:33:44.962] [DEBUG] StatusPoll(10.5.1.200): bAllDown=true, dynFlags=0x00000004
[06-May-2015 17:33:44.962] [DEBUG] Node::executeHookScript(10.5.1.200 [118]): hook script "Hook::StatusPoll" not found

This log is from NetXMS daemon running on -D 9 mode. It seems that there's a hook script missing.

[victor]

Missing script message is normal if you really do not have it (as most installations do). Hook script is optional and not required for normal system operation.

Best regards,
Victor

[jhuliagraziella]

Same  problem here.
I have a node that isn't accessible from my server and configurated an agent to poll it. I've tried remotely logging into the agent to ping this special node and everything worked fine.
The problem is: when I try to poll it from the server, even if I ask just for the status, it always stays as UNKNOWN. Any idea of what could be causing this?
I'm pretty sure my configuration is alright. The hostnames are correct and I've put the proxy in the agent.